title: Modeling Lexical Fields for Translation: a Corpus-Based Study of Armenian, German, and English Culinary Verbs
creator: Dallakyan, Meri
subject: ddc-400
subject: 400 Linguistics
subject: ddc-420
subject: 420 English
subject: ddc-430
subject: 430 Germanic
description: This doctoral thesis presents a corpus-based, bottom-up methodology for analyzing culinary lexical fields in Armenian, German, and English using statistical and computational tools combined with manual semantic annotation. Drawing from authentic comparable corpora, the study develops an interpretable model for identifying translation equivalents and typical lexical choices across languages, focusing on culinary verbs. The methodology offers practical applications for language learners, educators, translators, and developers of machine translation (MT) systems by minimizing atypical collocational usage and enhancing translation accuracy. A key innovation lies in the interpretability of the translation and lexical models, addressing a major shortcoming in current neural MT systems, which often lack transparency in their output. The proposed framework enables explainable lexical modeling and supports the evaluation and refinement of both dictionary definitions and MT outputs. While tested primarily on three languages, the model has demonstrated scalability, with preliminary success in French, and can be adapted to other languages with minor adjustments. By integrating manual semantic analysis with data-driven techniques, this work bridges the gap between logic-based and statistical approaches to language processing, offering a meaningful step toward explainable and semantically grounded applications in NLP and AI.
date: 2025
type: Dissertation
type: info:eu-repo/semantics/doctoralThesis
type: NonPeerReviewed
format: application/pdf
identifier: https://archiv.ub.uni-heidelberg.de/volltextserver/36999/1/Dissertation_Meri_Dallakyan.pdf
identifier: DOI:10.11588/heidok.00036999
identifier: urn:nbn:de:bsz:16-heidok-369991
identifier:   Dallakyan, Meri  (2025) Modeling Lexical Fields for Translation: a Corpus-Based Study of Armenian, German, and English Culinary Verbs.  [Dissertation]     
relation: https://archiv.ub.uni-heidelberg.de/volltextserver/36999/
rights: info:eu-repo/semantics/openAccess
rights: Please see front page of the work (Sorry, Dublin Core plugin does not recognise license id)
language: eng